Integrations

Feed Verax from any infrastructure

Verax is agentless by design: nothing to install, no traffic to decrypt, no inline appliance. If your stack can produce a list of destinations — and every proxy, DNS server, firewall, and SIEM can — you can be looking at your AI usage and its risk in minutes. One HTTP call.

Your workspace key

Each client gets a private, isolated workspace — your ingested data and findings are never mixed with anyone else's and never public. You send the key as an X-API-Key header on the calls below (the examples use $VERAX_KEY). We mint keys per pilot — no self-serve signup yet, so it stays a deliberate, scoped rollout.

Request a pilot key

Anything, as JSON

The universal path. POST rows with a destination hostname — field names are aliased (domain / host / hostname / query / url all work; same for user, src_ip, timestamp, bytes), so most exports work unmodified.

curl -X POST https://api.veraxai.xyz/ingest/generic \
  -H "X-API-Key: $VERAX_KEY" \
  -H "content-type: application/json" \
  -d '{"events": [
    {"destination": "chatgpt.com", "user": "jdoe", "src_ip": "10.1.2.3"},
    {"domain": "api.openai.com", "client": "10.1.2.9", "timestamp": 1751364000}
  ]}'

Anything, as CSV

Have a spreadsheet or a SIEM/proxy CSV export? Post it as-is — headers are matched by the same aliases. Unknown columns are ignored.

curl -X POST https://api.veraxai.xyz/ingest/generic/csv \
  -H "X-API-Key: $VERAX_KEY" \
  -H "content-type: text/csv" \
  --data-binary @export.csv

# export.csv
# date,host,username,src,bytes
# 2026-07-01T09:00:00Z,chatgpt.com,asmith,10.0.0.5,1200

DNS logs (Pi-hole, dnsmasq, Unbound…)

Every network has DNS — which makes it the fastest zero-install discovery source. Example: last 24h of Pi-hole queries, straight from its database.

sqlite3 -json /etc/pihole/pihole-FTL.db \
  "SELECT timestamp, domain, client FROM queries
   WHERE timestamp > strftime('%s','now','-1 day')" \
| jq '{events: .}' \
| curl -X POST https://api.veraxai.xyz/ingest/generic \
    -H "X-API-Key: $VERAX_KEY" \
    -H "content-type: application/json" --data-binary @-

Web proxy / secure gateway exports

Zscaler, Netskope, Squid, Bluecoat, Cloudflare Gateway — export web-transaction logs (CSV or JSON) and post them via the generic endpoints above. The columns those products emit (hostname, login, clientip, bytes_out…) are already aliased.

# e.g. Zscaler Web Insights export -> CSV -> one call
curl -X POST https://api.veraxai.xyz/ingest/generic/csv \
  -H "X-API-Key: $VERAX_KEY" \
  -H "content-type: text/csv" --data-binary @zscaler-export.csv

FortiGate logs (native parser)

Fortinet shops get a first-class parser for raw App-Control / web-filter log lines — no reshaping needed. Findings can be exported back to the fabric as CEF/syslog for FortiSIEM or FortiAnalyzer.

curl -X POST https://api.veraxai.xyz/ingest \
  -H "X-API-Key: $VERAX_KEY" \
  -H "content-type: application/json" \
  -d "{\"log\": \"$(cat fortigate.log)\"}"

# SIEM enrichment back out:
curl https://api.veraxai.xyz/export/cef      # CEF lines
curl https://api.veraxai.xyz/export/ocsf     # OCSF findings (JSON)

Microsoft Entra ID (OAuth grants)

Network logs can't see which AI apps hold standing OAuth access to your tenant. Export three read-only Graph collections and post them — Verax flags risky grants and generates the exact revoke commands.

az rest --url "https://graph.microsoft.com/v1.0/servicePrincipals?\$select=id,appId,displayName,publisherName"
az rest --url "https://graph.microsoft.com/v1.0/oauth2PermissionGrants"
az rest --url "https://graph.microsoft.com/v1.0/users?\$select=id,userPrincipalName,displayName"

# combine into one JSON object and:
curl -X POST https://api.veraxai.xyz/ingest/entra \
  -H "X-API-Key: $VERAX_KEY" -H "content-type: application/json" -d @entra.json

A different source?

If it can emit a hostname, it can feed Verax — Netflow, EDR telemetry, CASB exports, custom apps. Tell us what you run and we'll send the exact recipe, or build the parser.

Ask about your stack

See what the output looks like on the assessment console (demo data), or how facts are verified in the methodology.